Intro

This is a data set that includes the following variables:

DH1_W_Responsible
DH2_W_Civilized
DH3_W_Moral
DH4_W_Polite
DH5_W_Childlike
DH6_W_Rational
DH7_W_Warm
DH8_W_Agentic
DH9_W_Refined
DH10_W_Lacking_Culture
DH11_W_Lacking_Self-restraint
DH12_W_Instinctual
DH13_W_Mature
DH14_W_Stoic
DH15_W_Emotionally_Responsive
DH16_W_Cold
DH17_W_Open
DH18_W_Rigid
DH19_W_Passive
DH20_W_Superficial
DH1_N_Responsible
DH2_N_Civilized
DH3_N_Moral
DH4_N_Polite
DH5_N_Childlike
DH6_N_Rational
DH7_N_Warm
DH8_N_Agentic
DH9_N_Refined
DH10_N_Lacking_Culture
DH11_N_Lacking_Self-restraint
DH12_N_Instinctual
DH13_N_Mature
DH14_N_Stoic
DH15_N_Emotionally_Responsive
DH16_N_Cold
DH17_N_Open
DH18_N_Rigid
DH19_N_Passive
DH20_N_Superficial
No_Columbus
Create_IPD
BorninUS
LengthState
FatherUSBorn
MotherUSBorn
PrimaryCaretaker
PrimaryEd
SecondCaretaker
SecondEdu
Religion
Religiosity
College
Year
Education
LibCon
SES
Race
StateNumeric
LongestRegion
SubjectNumber
White_Dehumanize
Native_Dehumanize
Age


It’s primary focus is to look a bit more closely at the dehumaniztion variables and how they interact not only with one another, but with the dependent variables (Columbus Day/Indigenous Peoples’ Day (henceforth referred to as IPD)), as well as some of the demographic information gathered on the participants in this sample.

This sample comes from Study Two, so it is (if I’m not mistaken) a sample of participants gathered via mturk.

Descriptives

First, lets take a look at a descriptives table of our variables. Keep in mind that some of the variables are categorical and are more difficult to interpret in a descriptives table.

  n mean sd median se
DH1_W_Responsible 2811 61.24 21.82 62 0.4116
DH2_W_Civilized 2810 65.89 23.38 69 0.4411
DH3_W_Moral 2812 55.77 22.76 55 0.4292
DH4_W_Polite 2812 57.99 22.53 58 0.4248
DH5_W_Childlike 2809 40.76 24.49 43 0.462
DH6_W_Rational 2810 58.23 22.49 57 0.4243
DH7_W_Warm 2810 58.27 22.32 58 0.421
DH8_W_Agentic 2800 46.11 22.34 50 0.4222
DH9_W_Refined 2809 52.87 22.26 52 0.4199
DH10_W_Lacking_Culture 2808 46.34 27.71 50 0.523
DH11_W_Lacking_Self-restraint 2808 48.59 25.58 50 0.4828
DH12_W_Instinctual 2810 51.47 22.98 51 0.4335
DH13_W_Mature 2809 57.34 22.31 56 0.421
DH14_W_Stoic 2808 42.88 22.31 49 0.4209
DH15_W_Emotionally_Responsive 2812 59.98 22.77 60 0.4294
DH16_W_Cold 2810 44.74 23.99 49 0.4525
DH17_W_Open 2811 55.59 22.81 55 0.4302
DH18_W_Rigid 2807 49.6 23.27 50 0.4393
DH19_W_Passive 2810 45.16 23.21 49 0.4378
DH20_W_Superficial 2810 59.32 25.54 60 0.4818
DH1_N_Responsible 2807 63.86 21.06 64 0.3976
DH2_N_Civilized 2806 64.58 21.97 65 0.4147
DH3_N_Moral 2807 65.75 20.88 66 0.3941
DH4_N_Polite 2806 63.22 21.42 63 0.4044
DH5_N_Childlike 2802 29.22 22.26 25 0.4205
DH6_N_Rational 2807 61.61 21.17 61 0.3995
DH7_N_Warm 2804 61.05 22.07 60 0.4167
DH8_N_Agentic 2800 44.97 22.32 50 0.4219
DH9_N_Refined 2806 51.11 22.44 51 0.4235
DH10_N_Lacking_Culture 2804 23.9 23.15 17 0.4371
DH11_N_Lacking_Self-restraint 2805 34.44 23.39 33 0.4416
DH12_N_Instinctual 2807 61.3 23.48 61 0.4433
DH13_N_Mature 2806 65.15 20.69 65 0.3907
DH14_N_Stoic 2808 59.11 22.29 57 0.4206
DH15_N_Emotionally_Responsive 2806 54.38 24.29 53 0.4585
DH16_N_Cold 2805 35.19 22.8 35 0.4306
DH17_N_Open 2805 53.8 23.34 53 0.4406
DH18_N_Rigid 2807 45.41 23.89 50 0.451
DH19_N_Passive 2805 45.03 23.79 50 0.4492
DH20_N_Superficial 2805 30.41 22.87 27 0.4318
No_Columbus 2858 3.793 2.125 4 0.03976
Create_IPD 2858 4.481 1.956 4 0.03658
BorninUS* 2801 1.956 0.2041 2 0.003857
LengthState 2797 30.72 12.88 29 0.2435
FatherUSBorn* 2801 1.893 0.3352 2 0.006334
MotherUSBorn* 2801 1.882 0.3337 2 0.006306
PrimaryCaretaker* 2800 1.282 0.7601 1 0.01436
PrimaryEd 2801 2.988 1.225 3 0.02315
SecondCaretaker* 2801 2.828 1.114 3 0.02105
SecondEdu 2736 2.908 1.323 2 0.02529
Religion* 2799 2.45 1.674 2 0.03164
Religiosity 2800 3.632 2.235 4 0.04224
College* 2799 1.893 0.3094 2 0.005848
Year 300 3.363 1.631 3 0.09417
Education 2800 3.459 1.127 4 0.0213
LibCon 2799 4.401 1.713 4 0.03237
SES* 2801 1.538 0.4986 2 0.009422
Race* 2903 3.42 1.684 3 0.03125
StateNumeric* 2788 24.73 14.36 25 0.272
LongestRegion* 2789 4.425 2.104 5 0.03984
SubjectNumber 2903 1452 838.2 1452 15.56
White_Dehumanize 2793 52.93 12.03 53.25 0.2275
Native_Dehumanize 2792 50.67 11.38 51.05 0.2153
Age 2793 38.72 13.33 35 0.2523

Correlation of Dehumanization

One thing you’ll probably notice right away is that I’ve made some aggregate variables at the end of the data set that average across all Dehumanization measures for both whites and natives. Let’s take a look at the relationship between the two zoomed out a bit. 

## 
##  Pearson's product-moment correlation
## 
## data:  dfdehumanDV$White_Dehumanize and dfdehumanDV$Native_Dehumanize
## t = 55.589, df = 2784, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.7072142 0.7424398
## sample estimates:
##       cor 
## 0.7253014

At baseline the two seem to hang together pretty well. People seem to be rating Natives and Whites similarly among the measures of dehumanization that we have.

Dependent Variables: Columbus Day X Dehumanization

Before we move on, let’s take a gander at the dependent variables. Does there seem to be a relationship between dehumanization and Support for getting rid of Columbus Day?

## 
## Call:
## lm(formula = No_Columbus ~ Rating + Group + Rating:Group, data = dv4cor.lng)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.2443 -1.7991  0.1583  2.1348  4.5126 
## 
## Coefficients:
##                                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                    5.244311   0.180457  29.061  < 2e-16 ***
## Rating                        -0.027582   0.003325  -8.296  < 2e-16 ***
## GroupNative_Dehumanize        -1.296491   0.256706  -5.050 4.55e-07 ***
## Rating:GroupNative_Dehumanize  0.024315   0.004839   5.025 5.19e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.113 on 5581 degrees of freedom
##   (221 observations deleted due to missingness)
## Multiple R-squared:  0.01233,    Adjusted R-squared:  0.0118 
## F-statistic: 23.23 on 3 and 5581 DF,  p-value: 6.143e-15

**Keep in mind while looking at this graph that low scores on the dehumanization measure mean that participants are dehumanizing the group more.


Dependent Variables: Indigenous Peoples’ Day X Dehumanization

What about between dehumanization and Support for IPD?

## 
## Call:
## lm(formula = Create_IPD ~ Rating + Group + Rating:Group, data = dv5cor.lng)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.1305 -1.3811 -0.2253  1.5698  3.0665 
## 
## Coefficients:
##                                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                    5.130533   0.166784  30.762  < 2e-16 ***
## Rating                        -0.012417   0.003073  -4.041 5.40e-05 ***
## GroupNative_Dehumanize        -1.142140   0.237256  -4.814 1.52e-06 ***
## Rating:GroupNative_Dehumanize  0.021991   0.004472   4.917 9.03e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.953 on 5581 degrees of freedom
##   (221 observations deleted due to missingness)
## Multiple R-squared:  0.004461,   Adjusted R-squared:  0.003926 
## F-statistic: 8.337 on 3 and 5581 DF,  p-value: 1.577e-05

Dependent Variables: Side by Sides

We can also look at both of these groups side by side:


and…

Dependent Variables: Digging a Little Deeper

Whe can see more clearly that significant interaction when they are side by side. Attitudes toward white more strongly predicts the relationship between dehumanization and support for removing Columbus day.

Interestingly, the direction of that relationship changes when we’re talking about IPD. If we create a model that looks at this interaction with both ratings of support and dehumanization centered at the mean we see that this change in direction is significant (p = .005)

## 
## Call:
## lm(formula = Native_Dehumanize ~ Rating + Holiday_Support + Rating:Holiday_Support, 
##     data = dv4and5cor.lng.cen)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -51.489  -4.395   0.429   5.979  50.128 
## 
## Coefficients:
##                                   Estimate Std. Error t value Pr(>|t|)   
## (Intercept)                       -0.10887    0.21830  -0.499  0.61799   
## Rating                             0.32336    0.10994   2.941  0.00328 **
## Holiday_SupportNo_Columbus         0.07569    0.30859   0.245  0.80626   
## Rating:Holiday_SupportNo_Columbus -0.41695    0.14945  -2.790  0.00529 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 11.37 on 5580 degrees of freedom
##   (222 observations deleted due to missingness)
## Multiple R-squared:  0.001701,   Adjusted R-squared:  0.001164 
## F-statistic: 3.169 on 3 and 5580 DF,  p-value: 0.02334

**For some reason in the markdown document the "Rating" coefficient is significant. In the code it is not (p=.12).

Exploratory Analysis

Let’s dig into some of the demographic information in this subset of the data.

Race

First let’s look at the frequency of our different racial groups

Asian Black White Latino Middle Eastern Native Other Multiracial
139 224 2097 106 4 17 24 292

Since our data skews toward Asian, Black, White, and Latino let’s use these groups in future analysis.

Race X Support for Abolishing Columbus Day

First off, let’s look at how our different racial groups answered the question about Columbus Day.

Race X Support for Creating an Indigenous Peoples’ Day

Next let’s see how these groups answered the question about Indigenous Peoples’ Day

Race X Measures of Dehumanization: Whites

For the next set of graphs let’s revisit the question of dehumanization. What we’re interested in here is whether or not responses to the questions of dehumanization differed by racial group.

A little later on we’ll look specifically at questions around being civilized and childlike and whether these things predict responses to our DV questions and whether they interact with race at all.

Race N White_Dehumanize sd se ci
Asian 139 51.85 11.17 0.9476 1.874
Black 222 48.73 14.96 1.004 1.979
White 2082 53.57 11.54 0.2529 0.4959
Latino 105 53.65 12.13 1.183 2.347 

##               Df Sum Sq Mean Sq F value   Pr(>F)    
## Race           3   4952  1650.8    11.7 1.31e-07 ***
## Residuals   2544 359044   141.1                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 18 observations deleted due to missingness
##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = White_Dehumanize ~ Race, data = dfdehumanDV.race)
## 
## $Race
##                     diff        lwr       upr     p adj
## Black-Asian  -3.12266025 -6.4259093 0.1805888 0.0717501
## White-Asian   1.72137161 -0.9540863 4.3968295 0.3485664
## Latino-Asian  1.80477218 -2.1440217 5.7535661 0.6427737
## White-Black   4.84403186  2.6877973 7.0002664 0.0000001
## Latino-Black  4.92743243  1.3102173 8.5446476 0.0026465
## Latino-White  0.08340058 -2.9712461 3.1380473 0.9998760

This analysis is done via One way ANOVA using Dehumanization of Whites as the DV

Race X Measures of Dehumanization: Natives

Now let’s look at those same tables and bar graphs for Natives:

Race N Native_Dehumanize sd se ci
Asian 139 49.91 11.02 0.935 1.849
Black 220 50.31 11.92 0.8035 1.584
White 2084 50.61 11.22 0.2458 0.482
Latino 106 54.06 12.19 1.184 2.348 

##               Df Sum Sq Mean Sq F value Pr(>F)  
## Race           3   1332   443.9   3.468 0.0156 *
## Residuals   2545 325757   128.0                 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 17 observations deleted due to missingness
##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = Native_Dehumanize ~ Race, data = dfdehumanDV.race)
## 
## $Race
##                   diff        lwr      upr     p adj
## Black-Asian  0.3929971 -2.7582890 3.544283 0.9886075
## White-Asian  0.6937919 -1.8540521 3.241636 0.8970903
## Latino-Asian 4.1429585  0.3925199 7.893397 0.0235475
## White-Black  0.3007948 -1.7609760 2.362566 0.9820159
## Latino-Black 3.7499614  0.3111855 7.188737 0.0261964
## Latino-White 3.4491666  0.5532929 6.345040 0.0119136

This analysis is done via One way ANOVA using Dehumanization of Natives as the DV

Dehumanization and Holidays X Race: Dehumanizing Whites

Now, we’ll revisit the relationship between dehumanizing whites and support for the two holidays, this time broken down by race.

First Columbus Day:

Then Indigenous Peoples’ Day:

Dehumanization and Holidays X Race: Dehumanizing Natives

Now, onto the same analyses looking at Natives.

First Columbus Day:

Then Indigenous Peoples’ Day:

Stereotypes

Now we’re going to take the opportunity to drill down a bit into the question of Native Stereotypes.

This time we’re going to pull the stereotype of Natives as Childlike as well as the stereotype of Natives as Uncivilized.

The way I’m going to do this is to pull both measures out of the dehumanization composite score and look at their effects on our DVs separately. To test this statistically we’ll place them into linear models together with the composite.

Let’s begin!

Childlike

This first graph will show us regression lines for each of our 4 measures of dehumanization and their relationship with Support for abolishing Columbus Day

##                   vars    n    mean     sd  median trimmed     mad min
## SubjectNumber        1 2903 1452.00 838.17 1452.00 1452.00 1076.37   1
## No_Columbus          2 2858    3.79   2.13    4.00    3.74    2.97   1
## White_Dehumanize     3 2793   52.93  12.03   53.25   53.60    8.52   0
## Native_Dehumanize    4 2792   51.09  11.58   51.50   51.72    7.91   0
## DH5_W_Childlike      5 2809   40.76  24.49   43.00   39.98   26.69   0
## DH5_N_Childlike      6 2802   29.22  22.26   25.00   27.68   28.17   0
##                       max   range  skew kurtosis    se
## SubjectNumber     2903.00 2902.00  0.00    -1.20 15.56
## No_Columbus          7.00    6.00  0.17    -1.31  0.04
## White_Dehumanize    99.95   99.95 -0.87     3.45  0.23
## Native_Dehumanize  100.00  100.00 -0.87     4.02  0.22
## DH5_W_Childlike    100.00  100.00  0.21    -0.55  0.46
## DH5_N_Childlike    100.00  100.00  0.54    -0.43  0.42
## 
## Call:
## lm(formula = scale(No_Columbus, scale = F) ~ White_Dehumanize + 
##     Native_Dehumanize + DH5_W_Childlike + DH5_N_Childlike, data = dvstereotypes.child.c)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.8633 -1.8083 -0.0807  1.7581  4.6669 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        0.783234   0.188753   4.150 3.43e-05 ***
## White_Dehumanize  -0.053827   0.004709 -11.431  < 2e-16 ***
## Native_Dehumanize  0.027957   0.004910   5.694 1.37e-08 ***
## DH5_W_Childlike    0.018252   0.001726  10.575  < 2e-16 ***
## DH5_N_Childlike   -0.003985   0.001913  -2.082   0.0374 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.041 on 2781 degrees of freedom
##   (117 observations deleted due to missingness)
## Multiple R-squared:  0.07825,    Adjusted R-squared:  0.07693 
## F-statistic: 59.03 on 4 and 2781 DF,  p-value: < 2.2e-16

And we’ll do one for Indigenous Peoples’ Day as well:

##                   vars    n    mean     sd  median trimmed     mad min
## SubjectNumber        1 2903 1452.00 838.17 1452.00 1452.00 1076.37   1
## Create_IPD           2 2858    4.48   1.96    4.00    4.60    2.97   1
## White_Dehumanize     3 2793   52.93  12.03   53.25   53.60    8.52   0
## Native_Dehumanize    4 2792   51.09  11.58   51.50   51.72    7.91   0
## DH5_W_Childlike      5 2809   40.76  24.49   43.00   39.98   26.69   0
## DH5_N_Childlike      6 2802   29.22  22.26   25.00   27.68   28.17   0
##                       max   range  skew kurtosis    se
## SubjectNumber     2903.00 2902.00  0.00    -1.20 15.56
## Create_IPD           7.00    6.00 -0.35    -0.97  0.04
## White_Dehumanize    99.95   99.95 -0.87     3.45  0.23
## Native_Dehumanize  100.00  100.00 -0.87     4.02  0.22
## DH5_W_Childlike    100.00  100.00  0.21    -0.55  0.46
## DH5_N_Childlike    100.00  100.00  0.54    -0.43  0.42
## 
## Call:
## lm(formula = scale(Create_IPD, scale = F) ~ White_Dehumanize + 
##     Native_Dehumanize + DH5_W_Childlike + DH5_N_Childlike, data = dvstereotypes.child.i)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.5164 -1.3058  0.0292  1.6081  3.8440 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)       -0.033044   0.176129  -0.188    0.851    
## White_Dehumanize  -0.038655   0.004394  -8.798  < 2e-16 ***
## Native_Dehumanize  0.034383   0.004581   7.505 8.24e-14 ***
## DH5_W_Childlike    0.013216   0.001610   8.206 3.46e-16 ***
## DH5_N_Childlike   -0.007674   0.001785  -4.298 1.78e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.905 on 2781 degrees of freedom
##   (117 observations deleted due to missingness)
## Multiple R-squared:  0.05297,    Adjusted R-squared:  0.05161 
## F-statistic: 38.89 on 4 and 2781 DF,  p-value: < 2.2e-16

Civilized

Now we’ll take a look at stereotypes about being civilized using the same methods as before:

##                   vars    n    mean     sd  median trimmed     mad min
## SubjectNumber        1 2903 1452.00 838.17 1452.00 1452.00 1076.37   1
## No_Columbus          2 2858    3.79   2.13    4.00    3.74    2.97   1
## White_Dehumanize     3 2793   52.93  12.03   53.25   53.60    8.52   0
## Native_Dehumanize    4 2792   51.09  11.58   51.50   51.72    7.91   0
## DH2_W_Civilized      5 2810   65.89  23.38   69.00   67.81   25.20   0
## DH2_N_Civilized      6 2806   64.58  21.97   65.00   65.70   22.24   0
##                       max   range  skew kurtosis    se
## SubjectNumber     2903.00 2902.00  0.00    -1.20 15.56
## No_Columbus          7.00    6.00  0.17    -1.31  0.04
## White_Dehumanize    99.95   99.95 -0.87     3.45  0.23
## Native_Dehumanize  100.00  100.00 -0.87     4.02  0.22
## DH2_W_Civilized    100.00  100.00 -0.69     0.20  0.44
## DH2_N_Civilized    100.00  100.00 -0.48     0.15  0.41
## 
## Call:
## lm(formula = scale(No_Columbus, scale = F) ~ White_Dehumanize + 
##     Native_Dehumanize + DH2_W_Civilized + DH2_N_Civilized, data = dvstereotypes.civil.c)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.9888 -1.7404 -0.0827  1.6746  4.8252 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        0.949289   0.188085   5.047 4.77e-07 ***
## White_Dehumanize  -0.008293   0.005669  -1.463   0.1436    
## Native_Dehumanize  0.011076   0.005481   2.021   0.0434 *  
## DH2_W_Civilized   -0.028291   0.002188 -12.932  < 2e-16 ***
## DH2_N_Civilized    0.012068   0.002126   5.676 1.52e-08 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.019 on 2781 degrees of freedom
##   (117 observations deleted due to missingness)
## Multiple R-squared:  0.09807,    Adjusted R-squared:  0.09678 
## F-statistic:  75.6 on 4 and 2781 DF,  p-value: < 2.2e-16

And we’ll do one for Indigenous Peoples’ Day as well:

##                   vars    n    mean     sd  median trimmed     mad min
## SubjectNumber        1 2903 1452.00 838.17 1452.00 1452.00 1076.37   1
## Create_IPD           2 2858    4.48   1.96    4.00    4.60    2.97   1
## White_Dehumanize     3 2793   52.93  12.03   53.25   53.60    8.52   0
## Native_Dehumanize    4 2792   51.09  11.58   51.50   51.72    7.91   0
## DH2_W_Civilized      5 2810   65.89  23.38   69.00   67.81   25.20   0
## DH2_N_Civilized      6 2806   64.58  21.97   65.00   65.70   22.24   0
##                       max   range  skew kurtosis    se
## SubjectNumber     2903.00 2902.00  0.00    -1.20 15.56
## Create_IPD           7.00    6.00 -0.35    -0.97  0.04
## White_Dehumanize    99.95   99.95 -0.87     3.45  0.23
## Native_Dehumanize  100.00  100.00 -0.87     4.02  0.22
## DH2_W_Civilized    100.00  100.00 -0.69     0.20  0.44
## DH2_N_Civilized    100.00  100.00 -0.48     0.15  0.41
## 
## Call:
## lm(formula = scale(Create_IPD, scale = F) ~ White_Dehumanize + 
##     Native_Dehumanize + DH2_W_Civilized + DH2_N_Civilized, data = dvstereotypes.civil.i)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -5.530 -1.235  0.050  1.542  4.084 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        0.021525   0.176237   0.122   0.9028    
## White_Dehumanize  -0.009079   0.005312  -1.709   0.0875 .  
## Native_Dehumanize  0.015670   0.005136   3.051   0.0023 ** 
## DH2_W_Civilized   -0.018689   0.002050  -9.117  < 2e-16 ***
## DH2_N_Civilized    0.013673   0.001992   6.863 8.28e-12 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.892 on 2781 degrees of freedom
##   (117 observations deleted due to missingness)
## Multiple R-squared:  0.0656, Adjusted R-squared:  0.06425 
## F-statistic: 48.81 on 4 and 2781 DF,  p-value: < 2.2e-16

The story here seems to be that in the middle, when we’re talking about the average level of dehumanization, people are more or less treating the civilized question like they treat the other measures of dehumanization. However, when it comes to perceptions of Whites and Natives as childlike something else starts happening, particularly at the extremes. Whether there is a story there for These two measure in particular is difficult to say.

Childlike and Civilized by Race

Let’s look at these measures broken down by race and then by education level:

##                Df  Sum Sq Mean Sq  F value   Pr(>F)    
## Group           5 2597224  519445 1329.337  < 2e-16 ***
## Race            3   12148    4049   10.363 8.26e-07 ***
## Group:Race     15   50081    3339    8.544  < 2e-16 ***
## Residuals   15317 5985192     391                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 55 observations deleted due to missingness

And in case we haven’t done it, here’s the measure’s side by side without separating for race:

##                Df  Sum Sq Mean Sq F value Pr(>F)    
## Group           5 2597224  519445    1317 <2e-16 ***
## Residuals   15335 6047421     394                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 55 observations deleted due to missingness
##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = Rating ~ Group, data = dvstereotypes.race.lng)
## 
## $Group
##                                          diff        lwr         upr
## Native_Dehumanize-White_Dehumanize  -1.938569  -3.524086  -0.3530521
## DH2_W_Civilized-White_Dehumanize    13.523152  11.939801  15.1065030
## DH2_N_Civilized-White_Dehumanize    11.390845   9.807186  12.9745039
## DH5_W_Childlike-White_Dehumanize   -12.558589 -14.142093 -10.9750836
## DH5_N_Childlike-White_Dehumanize   -23.887754 -25.471876 -22.3036316
## DH2_W_Civilized-Native_Dehumanize   15.461721  13.878526  17.0449164
## DH2_N_Civilized-Native_Dehumanize   13.329414  11.745911  14.9129174
## DH5_W_Childlike-Native_Dehumanize  -10.620019 -12.203369  -9.0366701
## DH5_N_Childlike-Native_Dehumanize  -21.949185 -23.533151 -20.3652181
## DH2_N_Civilized-DH2_W_Civilized     -2.132307  -3.713642  -0.5509729
## DH5_W_Childlike-DH2_W_Civilized    -26.081741 -27.662921 -24.5005606
## DH5_N_Childlike-DH2_W_Civilized    -37.410906 -38.992704 -35.8291077
## DH5_W_Childlike-DH2_N_Civilized    -23.949433 -25.530922 -22.3679447
## DH5_N_Childlike-DH2_N_Civilized    -35.278599 -36.860705 -33.6964920
## DH5_N_Childlike-DH5_W_Childlike    -11.329165 -12.911118  -9.7472129
##                                        p adj
## Native_Dehumanize-White_Dehumanize 0.0065618
## DH2_W_Civilized-White_Dehumanize   0.0000000
## DH2_N_Civilized-White_Dehumanize   0.0000000
## DH5_W_Childlike-White_Dehumanize   0.0000000
## DH5_N_Childlike-White_Dehumanize   0.0000000
## DH2_W_Civilized-Native_Dehumanize  0.0000000
## DH2_N_Civilized-Native_Dehumanize  0.0000000
## DH5_W_Childlike-Native_Dehumanize  0.0000000
## DH5_N_Childlike-Native_Dehumanize  0.0000000
## DH2_N_Civilized-DH2_W_Civilized    0.0016998
## DH5_W_Childlike-DH2_W_Civilized    0.0000000
## DH5_N_Childlike-DH2_W_Civilized    0.0000000
## DH5_W_Childlike-DH2_N_Civilized    0.0000000
## DH5_N_Childlike-DH2_N_Civilized    0.0000000
## DH5_N_Childlike-DH5_W_Childlike    0.0000000

What we end up with is several possible stories. The big picture seems to be that in some cases it is the measure that is differentiating and other times the race of the participant seems to be driving differences, particularly in our civilized and childlike measures.

Childlike and Civilized by Education

Now, let’s take a bit of a left turn and introduce another wrinkle into this analysis. How do these things break down by education level. For the sake of argument we’ll look just at college educated versus those who are not college educated.

##                  Df  Sum Sq Mean Sq  F value   Pr(>F)    
## Group             5 2754957  550991 1378.219  < 2e-16 ***
## College           1    2858    2858    7.148  0.00751 ** 
## Group:College     5   21292    4258   10.652 3.09e-10 ***
## Residuals     16713 6681609     400                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 693 observations deleted due to missingness

A simple interaction plot will show us these effects in a different form:

Miscellaneous

This will be where we put other questions and answers that we have.

For now, #ByeFelicia